Jason tipped me off to the Strip Silence/Select All Forward techniques a few months after I started editing ATP, and it dramatically improved both my efficiency and the output quality.
Just be careful of the two big potential pitfalls to this approach:
Strip Silence serves effectively as a noise gate, so the same caution applies as noise gating: if there’s any audible hiss on the tracks, usually from full-blast MacBook fans or a noisy preamp with a high-gain mic, listeners will hear the noise cutting in and out between runs of speech. You’ll need to remove the hiss first with a software noise-profile remover so you’re left with tracks that are nearly perfectly silent when their people aren’t talking.
These all work similarly: you select a “silent” part of each speaker’s track that contains only the noise for a few seconds, capture that noise profile, then remove the profiled noise from the whole track. Audacity, Amadeus Pro, Sound Soap, and Adobe Audition all offer this. I’ve tried all four, and in my experience, the best one by far that leaves the fewest unnatural-sounding artifacts is Adobe Audition’s. (Of course, a better and probably cheaper solution is to eliminate the noise source.)
And when working with Strip Silence, leave some space around the newly chopped-up speech blocks so nobody sounds artificially cut-off too early. I like settings 1%, 1.2, 0.2, 0.3 in the Strip Silence dialog, which I’ve memorized because Logic doesn’t remember them between sessions and I need to enter them every time.